Mastercard expects more than 300 million online shoppers globally could routinely rely on artificial intelligence (AI) agents to shop and make payments on their behalf by 2030, signalling a potential shift in the architecture of digital commerce and payments.

The projection is contained in Mastercard’s report, A Short History of the Future of Shopping and Payments, that examines the implications of agentic AI for merchants, payment providers, banks, fintech companies and consumers.

The emerging model, known as agentic commerce, would allow AI systems to operate within parameters authorised by consumers, including searching for products, comparing prices and features, verifying information and ultimately initiating or completing transactions.

For the payments industry, the development could move AI from being primarily a customer-facing tool to becoming an active participant in the transaction chain.

AI agents could become new payment participants

Mastercard said routine purchases such as groceries, medicines and subscriptions are likely to be among the early categories in which consumers delegate purchasing activity to AI agents.

Under this model, a consumer could establish parameters such as budget, preferred brands, delivery requirements or other conditions, leaving the AI agent to determine which transaction meets those requirements.

The financial implication is that payment infrastructure would increasingly need to recognise and authenticate transactions initiated by software acting under a consumer’s authority.

This could create new requirements around identity, consent, authentication, transaction limits and liability, particularly where an AI agent has discretion to select a merchant or product.

Payment security moves beyond the consumer

Mastercard said trust will become increasingly important as software begins to transact on behalf of consumers.

The company identified identity, consent, authorisation, security and liability as areas that banks, fintech companies and merchants will need to address as agentic commerce develops.

The issue is significant for financial institutions because conventional digital payments generally assume that the person initiating a transaction is directly interacting with the merchant or payment interface.

Agentic commerce introduces another layer: software may initiate the transaction while the underlying financial authority remains with the consumer.

Payment networks and financial institutions could therefore face growing demand for mechanisms that distinguish authorised AI activity from unauthorised transactions, while giving consumers visibility and control over what their agents can purchase and spend.

Mastercard cites live European transaction

Mastercard said it and its partners have completed what the company described as Europe’s first live, end-to-end agentic payment transaction.

The company presented the transaction as a demonstration of how AI agents can participate in purchasing while maintaining consumer safeguards and control.

The development places payment networks at the centre of an emerging market in which the interface between consumers and merchants could increasingly be mediated by software.

For card networks and other payment infrastructure providers, the transition could create opportunities around authentication, tokenisation, transaction controls and other services required to support machine-initiated payments.

Retailers may have to optimise for machines

The shift could also affect the economics and operating models of digital commerce.

Mastercard said retailers will increasingly need to serve both human consumers and AI agents. Product information, claims, reviews, pricing, policies and other commercial data will need to be sufficiently structured and machine-readable for AI systems to evaluate them.

This means merchants may have to consider not only how products appear to consumers but also how effectively their commercial information can be discovered, interpreted and verified by AI systems.

The development could eventually influence digital advertising, product discovery, merchant visibility and customer acquisition as AI agents increasingly determine which products meet consumers’ requirements.

Implications for Nigeria’s payments market

For Nigeria, where banks, fintechs, payment service providers and merchants have rapidly expanded digital transaction channels, the development could introduce another layer to an already evolving payments ecosystem.

Agentic commerce could eventually require Nigerian payment providers to support controlled transactions initiated by AI systems, while maintaining clear links between the software, the authorised consumer and the underlying payment account.

It could also raise questions around transaction limits, dispute resolution, fraud liability, consumer protection and regulatory oversight when an AI system makes or executes a purchasing decision.

The opportunity extends beyond payments. Nigerian merchants seeking to participate in AI-mediated commerce may increasingly need to ensure that product catalogues, prices, inventory, delivery terms, refund policies and other commercial information are accessible to AI systems.

However, Mastercard’s 300 million figure is a projection rather than an established market size. Actual adoption will depend on consumer trust, regulatory developments, technical standards, merchant readiness and the ability of financial institutions and payment networks to provide secure infrastructure for machine-authorised transactions.

The broader significance for financial markets is that agentic AI could change not only how consumers shop, but also who, or what, initiates digital transactions, potentially creating a new layer of infrastructure between consumers, merchants and payment providers.